T/4 Fractionally Spaced Blind Equalizer And Its Extension Research

Different from the traditional adaptive equalization technology,blind equalization technology has the advantages of eliminating inter-symbol interference(ISI),and it does not need the training sequence,many scholars regard the blind equalization as a research subject.In a variety of blind equalization technology,fractionally spaced blind equalization technology can oversample the signal to get detailed channel parameters,at the same time,it can effectively avoid aliasing caused by undersampling and reduce steady state error.The thesis introduces the constant modulus algorithm,taking the structure of T/4 fractionally spaced equalizer as an example,because the constant modulus algorithm has slow convergence speed and large steady state error,the thesis studies a series of new theories and new technologies about T/4 fractionally spaced constant modulus blind equalizeation.The specific research contents are as follows:(1)This thesis introduces the basic theory of support vector machine.Meanwhile,it analyzes the mathematical theory of T/4-FSE-CMA-SVM.Finally,This thesis makes a simulation comparison between T/4-FSE-CMA-SVM and T/4-FSE-CMA,the results show that the T/4-FSE-CMA-SVM has fast convergence rate and small steady-state error.(2)Fractionally spaced blind equalization algorithm(FSE)has poor performance in balancing the channel of deep spectrum zero,and its convergence speed can be further improved,The thesis studies T/4 fractionally Spaced based on Momentum Constant Modulus Algorithm with Decision Feedback Equalizer,the decision feedback equalizer can effectively reduces the inter-symbol interference and the momentum algorithm can speed up the the algorithm of convergence speed.Finally,This thesis makes a simulateon comparison between T/4-FSE-CMA and T/4-FSE-MCMA-DFE,the results show that the T/4-FSE-CMA-SVM has fast convergence rate and small steady-state error.(3)This thesis analyzes a T/4 Fractionally Spaced Equalization algorithm based on Cosine Cost Function(T/4-FSE-CCF),at the same time,this thesis studies the simuliation of T/4-FSE-CCF.The core of T/4-FSE-CCF is to improve the cost function of the CMA and construct the cosine cost function,the T/4-FSE-CCF no longer depends on the statistical modulus R,Finally,this thesis makes a simulation comparison between T/2-FSE-CMA T/4-FSE-CMA and T/4-FSE-CCF,the results show that the T/4-FSE-C CF has fast convergence rate and small steady-state error.